A5101724374- Lung Cancer Diagnosis and Treatment
- Radiomics and Machine Learning in Medical Imaging
- Stochastic processes and financial applications
- Advanced X-ray and CT Imaging
- Cancer Diagnosis and Treatment
- Quantum Dots Synthesis And Properties
- Financial Risk and Volatility Modeling
- Financial Markets and Investment Strategies
- Electric and Hybrid Vehicle Technologies
- Cardiac Imaging and Diagnostics
- Molecular Junctions and Nanostructures
- Vehicle Dynamics and Control Systems
- Lung Cancer Treatments and Mutations
- ZnO doping and properties
- Electrocatalysts for Energy Conversion
- Hepatocellular Carcinoma Treatment and Prognosis
- Medical Imaging Techniques and Applications
- Capital Investment and Risk Analysis
- COVID-19 diagnosis using AI
- Cardiac Valve Diseases and Treatments
- Thin-Film Transistor Technologies
- Semiconductor materials and devices
- Risk and Portfolio Optimization
- Advanced Image Processing Techniques
- Electronic Packaging and Soldering Technologies
Seoul Women's University
2025
National Pension Service
2024
Electronics and Telecommunications Research Institute
2024
Asan Medical Center
2014-2023
Sungkyunkwan University
2011-2023
University of Ulsan
2014-2023
Ulsan College
2014-2023
Hanyang University
2020-2021
Kinokuniya
2019
Cambridge University Press
2019
Background Intratumor heterogeneity in lung cancer may influence outcomes. CT radiomics seeks to assess tumor features provide detailed imaging features. However, radiomic vary according the reconstruction kernel used for image generation. Purpose To investigate effect of different kernels on and whether conversion using a convolutional neural network (CNN) could improve reproducibility between kernels. Materials Methods In this retrospective analysis, patients underwent non–contrast...
This study is a retrospective survey of the variables that may influence development pneumothorax after thoracentesis. In 30-month period, computer search hospital records identified 342 thoracenteses, which 154 were done with conventional techniques by clinical services, and 188 sonographic guidance. Other factors surveyed included patients' age, sex, underlying pulmonary disease, overall condition; size effusion; type tap (diagnostic or therapeutic); amount (exudate transudate) fluid...
Objective:The aim of our study was to develop and validate a convolutional neural network (CNN) architecture convert CT images reconstructed with one kernel different reconstruction kernels without using sinogram.Materials Methods: This retrospective approved by the Institutional Review Board.Ten chest scans were performed B10f, B30f, B50f, B70f kernels.The dataset divided into six, two, two examinations for training, validation, testing, respectively.We constructed CNN consisting six...
Background: Asherman's syndrome (AS) is a dreadful gynecological disorder of the uterus characterized by intrauterine adhesion with severe fibrotic lesions, resulting in damaged basalis layer infertility.Despite extensive research on overcoming AS, evidence-based effective and reproducible treatments to improve structural functional morphology AS endometrium have not been established.Methods: Endometrial organoids generated from human or mouse endometrial tissues were transplanted into...
Emotion recognition using AI has garnered significant attention in recent years, particularly areas such as fashion, where understanding consumer sentiment can drive more personalized and effective marketing strategies. This study aims to propose an model that automatically analyzes the emotional emotions of fashion images compares performance CNN, ViT, ResNet determine most suitable model. The experimental results showed vision transformer (ViT) outperformed both ResNet50 CNN models. is due...
We estimate housing poverty determinants in the low-income middle-aged and elderly households using both Schwabe index 14th-18th Korea Welfare Panel Study data. Empirical findings are as follows: first, of 2022, 28.3% them were poverty. The biggest item on it was water heating cost, with without loans has increased consecutively lowest index, respectively. Second, little difference between single-family house apartment households, but that living other types, i.e, shanties, containers,...
Objective: To evaluate the accuracy of a deep learning-based automated segmentation left ventricle (LV) myocardium using cardiac CT. Materials and Methods: develop fully algorithm, 100 subjects with coronary artery disease were randomly selected as development set (50 training / 20 validation 30 internal test).An experienced radiologist generated manual set.The trained model was evaluated 1000 by an technician.Visual assessment performed to compare automatic segmentations.In quantitative...
Abstract We evaluated the efficacy of a curriculum learning strategy using two steps to detect pulmonary abnormalities including nodule[s], consolidation, interstitial opacity, pleural effusion, and pneumothorax with chest-PA X-ray (CXR) images from centres. CXR 6069 healthy subjects 3417 patients at AMC 1035 4404 SNUBH were obtained. Our approach involved steps. First, regional patterns thoracic identified by initial patch around abnormal lesions. Second, Resnet-50 was fine-tuned entire...
Abstract This paper examines a class of barrier options, multi‐step options , which can have any finite number barriers level. We obtain general, explicit expression for option prices this type under the Black–Scholes model by deriving reflection principle that is, boundary‐crossing probability Brownian motion. Multi‐step are not only useful in they handle different levels and time steps but also approximate with arbitrary barriers. Moreover, be applied to pricing jump‐diffusion models.
Intravascular ultrasound (IVUS) evaluation of coronary artery morphology is based on the lumen and vessel segmentation. This study aimed to develop an automatic segmentation algorithm validate performances for measuring quantitative IVUS parameters.
In this study, a co-operative regenerative braking control algorithm for the HEV with an electric booster brake was proposed and its performance evaluated. A dynamic model system derived to develop simulator investigated. addition, 6-speed AT-based hybrid vehicle established simulator, it combined simulator. Considering structure, strategy distribute force using booster. proposed, wherein torque is cut off ensure ride comfort maintain during downshift braking, then controlled considering...
This study measured parameters automatically by marking the point for measuring each parameter on whole-spine radiographs. Between January 2020 and December 2021, 1017 sequential lateral radiographs were retrospectively obtained. Of these, 819 198 used training testing performance of landmark detection model, respectively. To objectively evaluate program’s performance, 690 from four other institutions external validation. The combined dataset comprised 857 female 850 male patients (average...
Abstract We investigated whether a convolutional neural network (CNN) can enhance the usability of computer‐aided detection (CAD) chest radiographs for various pulmonary abnormal lesions. The numbers normal and patients were 6055 3463, respectively. Two radiologists delineated regions interest lesions labeled disease types as ground truths. datasets split into training, tuning, testing 7:1: 2. Total test sets randomly selected in 1214 690 abnormal. A 5‐fold, cross‐validation was performed on...
To understand the steric effects on formation of alicyclic thiolate self‐assembled monolayers Au(111), surface structures and reductive desorption behaviors Au(111) derived from cyclohexanethiol (CHT) 1‐methylcyclohexanethiols (MCHT) were characterized using scanning tunneling microscopy (STM) cyclic voltammetry. STM observations showed that adsorption CHT resulted in well‐organized with an unit cell a = 14.5 ± 0.3 Å b 10.2 Å, while MCHT methyl substituent bonded to 1‐position cyclohexyl...
In this study, a regenerative braking control algorithm was developed for the fuel cell electric vehicle (FCEV) equipped with front electronic wedge brake (EWB). A proposed and friction wheels, rear wheels; powertrain model of an FCEV composed motor, cells, batteries, MATLAB algorithm, were developed. They linked to SimDriveline under whereas hardware in loop simulator (HILS) built using EWB hardware. HILS used examine dynamic response motor torque clamping force that corresponded driver's demand.
현대 기업경영 패러다임이 주주 자본주의에서 이해관계자 자본주의로 전환되면서 환경․사회․지배구조(ESG)경영이 기업의 사회적 책임의 필수적인 부분으로 부각되고 있다. 소비자들의 기업에 대한 관심이 증가함에 따라 ESG활동이 소비자의 구매에 미치는 영향은 점점 더 커지고 있으며, 이로 인해 ESG활동을 마케팅 PR(MPR) 수단으로 활용하는 사례도 증가하고 그러나 지금까지의 연구는 재무 및 성과에 영향에 중점을 두어왔을 뿐, 기업이 가치를 높이기 위해 어떻게 효과적으로 설계하고 실행할지에 부족하였다. 이에 본 연구에서는 주요 ESG경영활동을 바탕으로 컨조인트 분석(conjoint analysis)을 수행하여, 고객의 요구와 선호를 반영한 최적의 ESG경영활동 설계방안에 대해 연구하였다.BR 연구의 분석결과는 다음과 같다. 분석을 수행한 결과, 응답자들은 환경(E) 활동 중 에너지효율화조치를 가장 중요하게 생각하였으며, 생물다양성 보존, 기후변화전략, 자원관리 등이 그 뒤를...
There are many technical issues regarding chiplet integration. For example, there concerns about how to obtain clean and reliable solder cap surfaces, Cu RDL pillars below 40μm pitch, achieve precise bonding accuracy during the Chip-on-Wafer (CoW) process.In this paper, we propose a room temperature homogenized laser with water-soluble epoxy flux as candidate solution first issue, Laser-Assisted Bonding Compression (LABC) Non-Conductive Film (NCF) for second problem, compare it other...
Growth processes and electrochemical behaviors of 4-fluorobenzenethiol (4-FBT) self-assembled monolayers (SAMs) on Au(111) prepared by vapor deposition at 323 K were examined using scanning tunneling microscopy (STM) cyclic voltammetry (CV). STM imaging revealed that 4-FBT SAMs the initial growth stage (deposition for 1 min) mainly composed bright molecular aggregates liquid-like disordered phase. After longer 3 min, had three distinct surface features: a few aggregates, small ordered...